Transaction Cost and the Viability of Rural Financial Intermediaries

Number Twenty Eight, Volume XV I, No, I,1989

TRANSACTION COST AND THE VIABILITY OF RURAL FINANCIAL INTERMEDIARIES* Teodoro S. Unta/anand Carlos E. Cuevas**

Introduction Rationaleand Objectivesof the Study The definition of transaction cost asused here departs from its usual meaning which is the measure of ffriction" in any intermediation process.

Transaction cost in this study is the cost incurred as banks perform the role of intermediator among savers and user of funds. This results from bank operations in lendingand mobilizingfunds, as well as from other operations, e.g., investments. Transaction cost includes administrative costs, i,e., personnel and fixed costs as well as risk-related costs, that are normally encounteredin dispensing and keepingthesefunds. It isa vital aspect ofthe bank's operational capability and largely determines the bank's viability as an intermediary.

The process of intermediation is the result of banks' comparative advantage in bringing about a market mechanism for the efficient transfer of claims on resources from surplus units to deficit units. High transaction cost runs against this rationale and impedes the intermediary's efficiency in resource allocation and distribution.

Against the backdrop of the increasing need to provide credit to the agricultural sector, the continued existence of intermediaries in the rural sector is necessary. The present thrust of Philippine economic development of uplifting the income of rural families through the growth of the

*Paper presented during the ACPC-PIDS-OSU sponsored seminar-workshop on "Financial Intermediation in the Rural Sector: Research Results and Policy Issues"held on 26-27 September 1988at the Cuaderno Hall, Central Bank of the Philippines. This is part of a larger study on comparative bank analysis jointly conducted by the Agricultural Credit PolicyCouncil (ACPC), PhilippineInstitutefor DevelopmentStudies(PIDS), and OhioState University (OSU). The project was coordinated by Dr. Mario B. Lamberte(PIDS) and Dr. V. Bruce J. Tolentino (ACPC). **Respectively, Research Associate, PIDS and Professor,OSU. The views expressed in this study are those of the authors and do not necessarily reflect those of the Institute,

38

JOURNAL

OF PHILIPPINE

DEVELOPMENT

agricultural sector only serves to highlight the need for a continuous supply of credit to rural households. Over the years, the share of agricultural credit to total credit has substantially declined and yet agriculture has remained to be a very important source of livelihood for most Filipinos considering that about 81 percent of Filipino families in the lowest 30 per(_ent income class derive their income from agriculture (Tolentino 1987).

In the past, attempts were made to infuse cheap funds into the rural sector through the formal financial system with the hope that the availability of credit could stimulate the development of the agricultural sector. While the intention of providing cheap credit is noble, it overlooks its adverse effects on the transaction cost of banks. Banks' cost of administering donor-sourced funds could be high, thus affecting their operations and compromising their viability (Cuevas 1984). The recognition of this problem has recently led to some policy changes. Apart from ensuring the continuous flow of credit to the rural sector, the new set of policies also seeks to protect banks from incurring unnecessarily high intermediation costs.

This paper attempts to examine the transaction cost of banks. Its specific objectives are: 1) to develop a method of estimating transaction cost for each bank activity, i.e., lending cost, funds mobilization and general administration; 2) to explain the differences and the composition of transaction cost among commercial banks (KBs), private development banks (PDBs), and rural banks (RBs).

Recognizing the need to continually introduce improvements for the efficient functioning of the formal financial system as a sector vital to economic growth, knowledge of banks' transaction cost is important. It can serve as a policy benchmark onwhich future changes and improvements in the financial system can be based. These may in turn induce banks to assume a wider role in the whole financial process ensuring a stable flow of credit to the rural sector.

Organization

of the Study

section II discusses the components of bank cost. A detailed presentation of the methods and procedures of estimating the transaction cost of banks and a description of the sources and limitations of data are given in Section II1. Section IV presents the empirical findings of the study. Finally, Section V summarizes the results and discusses some policy implications.

UNTALANandCUEVAS:

Conceptual

TRANSACTION

COST

39

Issues

A bank incurs costs in the process of mobilizingand lendingfunds.

These costs may be grouped into three categories. First, the interest cost paid to its depositors or its interest cost. Second, the incidental expensesincurred such as insurance for itsdeposits, insurance premia for its loans, as well as fines and penalties. Finally, the administrative costs such as the salaries and depreciation cost to bank premises, furniture, fixtures, equipments and others.

These costs, together with the interest cost of funds, determine the overall costs or total cost of intermediation for a bank: The bank's transaction cost can be summarized as: TCost = LCost + FCost + GCost + OCost where TCost Lcost FCost GCost OCost

Cost

= transaction cost of the bank = lendingcost = funds-mobilization cost = general administration cost = other operationalcosts,e.g. investments

of Funds

The bank incurs financial expenses in the form of interest payments paid to depositors. Similarly, the bank pays interest on funds obtained from the Central Bank rediscounting window, borrowings from other banks, and/or special lending programs. These are the bank's pure costs of funds or interest costs.

Incidental

Expenses

In its lendingoperations,the bank incursrisk-relatedcostsforits loan deliveryand recovery. These maycome in the formof guaranteefeesor insurancepremia applicable to particular loans in the bank's portfolio when it participateswiththe special lendingprograms. These are necessarycosts for banks as additionalsecurityagainstdefaultsor bad debts.

In addition,banks also incurcostsin the provisionsfor bad debts,or loan defaultas well a litigationexpensesassociatedwith the foreclosures.

The bank also incursincidentalexpenses in its funds-mobilization

40

JOURNAL

OF PHILIPPINE

DEVELOPMENT

such as deposit insurance, and the fines and penalties paid by the bank when it cannot meet the reserve requirements.

Administrative

Costs

In performing its funds-mobilization and lending operations, banks incur variable and fixed expenses. On the funds- mobilization side, variable expenses correspond to the salaries paid to personnel involved in the bank's deposit-taking and borrowing operations. Fixed costs associated with funds mobilization are its share of depreciation costs on building, fixtures and equipment used in the bank's operation and other overhead expenses.

Similarly, administrative expenses such as salaries for personnel involved in loan processing, supervision, monitoring, and collection activities are incurred in the bank's lending operations. These also have their counterpart in the depreciation costs of the building, fixtures and equipment as well as overhead expenses.

There are administrative costs of the bank which are clearly identifiable in terms of lending operations and deposit-mobilization activities.

Where other costs cannot be directly or specifically associated with any of the bank's major activities, then these are considered as general administration costs which are incurred in other operations of the bank. Inthe same manner as funds-mobilization and lending costs, these include salaries for personnel involved in general administrative work and depreciation cost for the building, fixtures and equipment and other expenses related to such operations.

Opportunity

Cost of Funds

Imputed costs of funds result from the opportunity cost of funds locked in loans overdue. Similarly, opportunity costs may be imputed by some banks due to the differences in the required reserves for these banks. Computation may be based on the market cost of funds applied to the total volume of funds under consideration. However, this cost is not considered in this study.

Methods

and Procedures

This sectionpresentsthe methodsand proceduresused in estimating thetransactioncostof banksfromthe set of primarydata.

UNTALAN

and CUEVAS:

Time Allocation

TRANSACTION

COST

and Transaction

41

Cost

A table of time-allocation for the different functions in a bank was completed by each bank staff (Table l-A). Each staff was asked to give, in percentages, the time allocated to each of the pre-identified bank activities (Appendix 1). A corresponding personnel compensation table was completed with the monthly salary for each bank staff (Table l-B).

From these initial data, estimation of the values of transaction cost for each bank is done by first, giving weights to the percentage of time allocated by each personnel by using the salary of that personnel from the personnel compensation table. This is done by multiplying the time allocationof one personnel for the different bank activities (Table 1-A) by his corresponding salary (Table l-B).

In order to reduce the number of variables needed in processing the data, bank personnel or positions were grouped into classes having the same or similar functions and were assigned one variable (Appendix 2).

The guidelines followed are given in Appendix 3.

From the weighted percentages of time-allocation provided by each bank personnel for the different bank functions, a horizontal summation for each function across all bank personnel was made. Using this weighted time-allocation for each bank activity, the percent share of each function was taken from the total (Table 2-A).

Columns (1), (2), and (3) show the weighted time allocation in pesos for the different positions.

The weighted time allocation for each bank activity across all bank personnel is summed up in column (4). The percentages in column (5) are then derived by taking the share of each bank activity to the total as given in column (4).

The resulting shares in percentages were used to allocate personnel costs i.e., salaries, and non-personnel costs, i.e., depreciation trom the bank's income and expense statements (1 able 2-B). i-or other expense items appearing in the income and expense statements which are clearly identifiable with specific bank activities, e.g., deposit insurance or guarantee fees, these are immediately allocated to that particular bank activity (Table 3). Column (1) gives the share of each bank activity in the total expenses on personnel, i.e., salaries, benefits, bonuses for each bank activity. In this case, column 1, item A is the share of the bank's lending operations in the total expenses on bank personnel. In the same manner, the share of each bank activity in the bank's non-personnel expenses, i.e., depreciation, taxes are given in column (3). Column (2) is the direct allocation of costs specific of a bank activity. For example, insurance premia is directly attributed to lending cost since these are costs related to lending. The sum of both personnel and non-personnel costs

42

JOURNAL

OF PHILIPPINE

DEVELOPMENT

Table 1 A, TIME-ALLOCATION TABLE

Bank Activity/ Function •

Manager

A. B. C. D.

Lending(1-10) 50% Investments Trust FundsMobilization(1 - 5) 50% E. Gen. Administration TOTAL

100%

Bank Personnel Accountant Teller 70%

30%

100%

100%

100%

B. PERSONNEL COMPENSATION TABLE 1. Manager

P1,000

2. Accountant

P 700

3. Teller

P 500

* See Appendix 1 for breakdown for each bank activityis given in column (4). The correspondingshares of each bankactivityfromthe totalis givenin column(5).

The allocationor breakdownof the transactioncost for each major bank activitycan then be derived by takingthe shares of these activities in totalcosts.

Sources and Description of Data This study used a sample of 64 out of a total of 66 banks classified accordingto type, location,and class. Two of the totalnumberof banks did notgiveany informationrelatedto the aspectswhichwere considered in this study. Of this sample,22 are ruralbanks, 17 are privatedevelopmentbanks, and 25 are commercialbanks.All of the bankssampledare locatedoutsideMetro-Manilaor are consideredas operatingin a rural or

UNTALAN

and CUEVAS:

TRANSACTION

COST

43

Table 2 A. WEIGHTED TIME-ALLOCATION TABLE

Bank Personnel Bank Activity/ Function

Manager Accountant Teller

(1) a. b. c. d.

Lending Investment Trust FundsMobilization e. Gen. Administration

TOTAL B.

Total

"P500

(2)

(4)

(5)

P990

45%

1='500 'P1210

55%

=PS00 "P2200

100%

t=490

=1=500 =P210

"P1000

(3)

1_'/00

BANK'S INCOME AND EXPENSE STATEMENTS

a. Salariesand Wages (Personnel) b, Depreciation/ Other OperatingExpenses (Non-Personnel)

sub-urbansetting.

The data gatheredwere qualitativeand quantitativeresponsesto the survey questionnaireaugmentedby supportingdocumentssuch as income statements, balance sheets, and job descriptions.The data was from a singleyear fromJanuaryto December 1986.

Raw data were obtainedon the time-allocationof each personnelfor the different functionsof a bank. Each bank personnelwas represented as everyone is madeto respondto thetime-allocationtable.The basicinformationobtainedwas the percentageof the time of each personnelallocatedper function.

44

JOURNAL

OF PHILIPPINE

DEVELOPMENT

Table 3 TRANSACTION COST

Bank Activity/ Function

Personnel Cost

(1) A. Lending

_

B. Investment

Non-Personnel Cost

Total

Exclusive Non-exclusive P (2) (3) (4)

% (5)

insurance premia

=ls=

A

%

:P=

=1 ==

B

%

C. Trust

=t=

=1 ==

C

%

D. Funds Mobilization E. Gen. Admin.

=1 =

=P=

D

%

=P'

E

%

deposit insurance

t==

r

Transaction Cost =

100%

Thus, from the foregoing: Transaction Cost (TCost) = A ,+B + C + D + E, where A B C D E

is the total cost of lending by the bank; is the total cost for investment operations; is the total cost for trust operations; is the total cost for funds-mobilization; and is the total cost for general administration.

Limitations of the Data

Quantitativeresponsesfrom the set of primary data collected maybe partly qualitative in nature as these may depend on the respondent's interpretation of the question at hand and the time-frame. This maybe particularly true where the respondents were asked about the allocation of

UNTALANandCUEVAS:

TRANSACTION

COST

45

their time to the different bank activities. Nevertheless, their responses maybe considered as best estimates.

In addition, answer to such questions as loan as a percentage of collateral, and number of repeat borrowers serviced were based on best estimates by the respondent in cases where bank records were not readily available.

Lastly, data on bank expenses although lifted straight from the banks' income and expense statements, may not exactly reflect actual costs incurred for some bank activities. This is particularly true for KBs and PDBs where loan processing costs may be undervalued since part of activities of processing a loan are done at the head office but these costs may not be properly accounted for by the branch. These might have produced biased estimates.

Empirical

Results

This section discusses several sets of results. The first sub-section provides an overview of the transactions costs and the different composition of these costs among KBs, PDBs, and RBs. The next sub-sections, pp. 55 and 64, focus on the lending costs and funds-mobilization costs, respectively, among the different bank types. These sub-sections present the composition of the costs of lending and funds-mobilization, two of the banks major operations, and attempt to explain the differences in this composition across the three bank types.

Sub-sections on pp. 66 and 73 relate the costs of lending and funds mobilization to the respective number and value of loans and deposits, in order to determine the per unit cost of providing these services. The costs per loan and per deposit provide indication of the comparative advantage of different bank types in providing these services to their customers.

Total Transaction

Cost: A Comparative

Analysis

Forty-nineof the 66 banks and bank branches in the sample reported time-allocation tables and income statements. Of these 49 banks, 16 are rural banks, 14 are private development banks and 19 are commercial banks.

Total transaction cost for the overall sample of forty nine banks combined are due primarily to funds-mobilization activities, 49.8% (see Table 4). Transaction cost associated with lending operations account only for 27.9 percent of the total. Bank activities related to administrative and general services account for 20.9 percent of the total transaction

Table 4 TRANSACTION COST (in thousand pesos) ALL BANKS Pesos TOTAL LENDI NG COSTS:

KBs %

Pesos

%

RBs

Pesos

%

Pesos

38.55

%

20694,47

27.85

19.72

5401.50

5789,64

47.83

1488.17

1.98

902.52,

1,87

233,83

1.67

331.82

2,74

Ads & Promo

440.33

0.59

!62.58

0.34

115,65

0.83

162.10

1,34

Disbursement

1054,60

1.42

412.11

0.86

288,98

2.05

355.51

2.94

Unspecified

2886,99

3.89

1128.10

2.34

677,85

4.84

1081.04

8.93

Loan Processing: Interview of Applicants

1646,16

2,22

693.64

1.44

446.38

3.19

505.94

4.18

Credit Investigation

2356,23

3,17

1186.04

2.46

709.32

5.06

460.87

3,81

Evaluation & Analysis

2305,38

3,10

1229.27

2.55

382.41

2.73

693.70

5.73

6307.760

8.49

3109,15

6.45

1538.11

10.98

1660.50

13,72

Planning & Programming

9503.33

PDBs

c _0 z o -n I r-

Loan Recovery: Monitoring Collection

-_ "o

f 054.95 1526.89

1,42 2.06

641,08 518.33

1.33 1.08

212.43 560.96

1.52 4.00

201.44 447.60

1.66 3.70

Record-keeping/

rn o m <

Report-writing •

2886.19

3,88

1148.54

2.38

880.90

6.29

856.75

7.08

m f-

Mgt. of bad debts

3068.59

4.13

1480.91

3.07

894.79

6.39

692.89

5.72

o

8536.63

11.49

3788.86

7.86

2549.08

16.19

2198.68

18.16

£rl

z -I

c z .H 3> i-

Table 4, continuation

> z ALL BANKS Pesos

KBs %

Pesos

PDBs %

Pesos

RBs %

Pesos

%

cL tm < _>

Cq

INVESTMENTS

569.19

0.77

83.71

0,17

272,89

1.95

212,59

1.76

TRUST OPERATIONS

522.72

0.70

402.09

0.83

112.18

0.80

8.46

0.07

:0 _> z Gq

TOTAL FUNDSMOBILIZATION

> COSTS:

37010.21

49.81

27241.05

56.53

5146.21

36.73

4622.95

38.19

2335.82

3.14

1250.52

2.60

274.40

1.96

810.90

6.70

17636.44

23.74

14223.77

29.52

2250.75

16.07

1161.92

9.60

z L-3 O tJ3

Record-keeping

8589.63

11.56

6443.39

13.37

1293.56

9.23

852.68

7.04

-.I

Funds-Transfer

1529.69

2.06

1098.31

2.28

355,90

2.54

75.48

0.62

Ads & Promo

2250.03

3.03

1552.76

3.22

380.58

2.72

316.70

2.62

Unspecified

4663.80

6.28

2671.77

5.54

586,75

4.19

1405.27

11.61

15502.65

20.87

10954.44

22.73

3077.32

21.97

1470.88

t 2.15

74299.23

100.00

48184.61

100.00

14010.1

100.00

12104.51

100.00

Transactions with CB, other banks Transactions with Depositors

-I

6

GEN ERAL ADMIN ISTRATiON/ SERVICES

TRANSACTIONS

COSTS:

Source: Comparative Bank Study, 1987 .1=, ,,,.,i

48

JOURNAL

OF PHILIPPINE

DEVELOPMENT

cost, while the rest corresponds to other bank operations such as investment, o.77 percent, and trust, 0.70 percent.

Transactions with bank depositors and clients represent almost onehalf of the costs of funds-mobilization activities for these banks. This indicates that an important proportion of bank resources is allocated to raising funds from the public for their operations. Activities related to transactions between these banks and the Central Bank (CB) and other banks account for only 3.1 percent of total transaction cost. The disparity, in these shares in costs underlines the preference by these banks to source their funds from the public rather than from other sources. However, the cost of mobilizing funds from the Central Bank is not negligible, as is usually assumed.

In their lending operations, activities related to ioan processing account for 8.5 percent of total transaction cost. A larger proportion (11.5%) of their costs is attributed to loan recovery efforts. The latter suggests a cautious attitude towards lending and the banks' greater effort to recover funds. It is noteworthy that banks incur minimal costs in promoting its lending activities. This suggests that banks do not really exert effort to attract prospective borrowers. Banks, on the other hand, incur higher costs relative to their total transaction cost in their deposit mobilization activities, indicating that banks make a more serious effort in attracting depositors than borrowers. This is shown by the relative shares in total transaction cost of advertising and promotions cost specific to loans (0.6%) against advertising and promotions cost specific to deposits (3.0%).

Among types of banks, transaction cost on the average are highest among KBs, P2.5M, followed by.PDBs, P1M, with RBs having the lowest transaction cost, P.79M, (see Table 5). There is a greater dispersion in transaction cost among KBs, followed by PDBs then RBs. About 36.8 percent of the commercial banks surveyed have transaction cost above P2.5M. Most PDBs and RBs have transaction cost of P1M or less, respectively. This is to be expected since KBs have bigger operations than PDBs and RBs. They have more personnel allocated to provide various services to their clientele. A typical KB has on the average a staff of 22 with PDBs having 16 and RBs 16 including officers and management personnel (Table 6). KBs also have higher fixed costs, i.e., depreciation for their building and equipments. Another important contributing factor is the relatively higher salary scale of KB personnel than either PDBs and RBs. This only serves to underscore the fact that the size of the bank has a bearing on the magnitude of its transaction cost.

When the composition of transaction cost is compared among bank types (Table 4), more than half (56.5%) of overall transaction cost of KBs come from funds-mobilization. Only 19.7 percent of their transaction cost

Table 5 • COMPARATIVETRANSACTION COST (in thousand pesos) KBs

PDBs

r> z KBs

PDBs

RBs

%

%

%

0.00

50.00

ROW Number

500 & Jess

_>

RBs

"F_NSACTIONS COST

c ;_

%

Number

%

0

0.00

3

21,43

1000 & less

1

5.26

6

1500 & less

1

5.26

2

2000 & less

4

21.05

2500 & less

6

31.58

2500 +

7 19

Number

%

cm

TOTAL

50.00

,< _>

3

18.76

6

42,86

8

50.00

15

6.67

40,00

53.33

14.29

5

31.25

8

12.50

25.00

62,50

2

14,29

0

0,00

6

66.67

33,33

0.00

-_ -n _> z _, c_ "_

0

0,00

0

0,00

6

100.00

0.00

0,00

z

36.84

1

7,14

0

0.00

8

87.50

12,50

0.00

100

14

100

16

100

C)

TOTAL

AVG:

2536.032

1000.721

798,582

SD:

1020,332

640,038

345.788

VAR:

1041078

409649

119569

Source:

O

Comparative Bank Study. 1987

4_ t,D

ol

Table 6 PERSONNEL DISTRIBUTION KBs

PDBs

RBs

KBs

PDBs

RBs

Row PCode*

index

Number

%

Humber

%

Humber

%

Total

% Total

%

%

%r

1 2

0 0

0,00 0.00

5 5

'1,98 1.98

12 10

4,38 3,65

17 15

1,57 1.38

0.00 0.00

29.41 33,33

70.59 66.67

3 4

0 0

0.00 0.00

20 5

7,91 1,98

50 12

18.25 4.38

70 17

6.45 1,57

0.00 0.00

28.57 29.41

71.43 70,59

5 6

24 17

4.30 3,05

16 5

6.32 1.98

16 4

5.84 1.46

56 26

5. I 6 2.40

42.86 65.38

28.57 19,23

28.57 15.38

7 8

27 17

4.84 3.05

15 18

5, 93 7.11

30 20

10.95 7.30

72 55

6.64 5.07

37.50 30,91

20.83 32.73

41.67 36,36

9 10

L F

5 12t

0.90 21,68

6 33

2.37 13.04

6 12

2.19 4.38

17 166

1.57 15.30

29.41 72.89

35.29 19.88

35.29 7.23

c ::0

11 12

F F

28 63

5.02 11.29

1 21

0,40 8.30

2 16

0.73 5.84

31 1O0

2.86 9.22

90.32 63.00

3.23 2t .00

6.45 16,00

r-_>

13

F

24

4,30

9

3.56

0

0.00

33

3,04

72.73

27.27

0,00

14 15

L L

20 12

3.58 2.15

16 13

6.32 5,14

21 14

7.66 5.11

57 39

5.25 3.59

35.09 - 30.77

28.07 33.33

36.84 35.90

16 17

L F

0 11

0.00 1,97

I 3

0.40 1.19

14 0

5.11 0.00

15 14

1.38 1,29

0.00 78.57

6.67 21,43

93.33 0.00

.'O._

2 178

0.36 31 .gO

1 48

0.40 18.97

5 19

1.82 6.93

8 245

0.74 22.56

25,00 72.65

12.50 19.59

62.50 7.76

m om <

18 19

0 -n..o 'r-

m r-

0

'1o m

z .-I

c z

Table 6, contlnuatlon

-_ 3>

KBs PCode*

index

PDBs

Number

%

Number

RRs %

Number

KBs %

Row Total

% Total

%

PDBs %

RBs %

gz

20

L

2

0,36

4

1.58

5

1.82

11

1.01

18.18

36.36

45.45

c m < 3>

21

L

0

0.00

5

1.98

4

1.46

9

0.83

0.00

55.56

44.44

""

L

0 0

0.00 0,00

0 1

0.00 0,40

1 0

0.36 0,00

1 1

0.09 0.09

0.00 0,00

0.00 100.0

100.00 0,00

24

1

0.18

2

0,79

1

0.36

4

0.37

25.00

50.00

25.00

25 26

6 0

1.08 0.00

0 0

0.00 0.00

0 0

0.00 0.00

6 0

0.55 0.00

10000 0.00

22 23

0.00 0.00

O'3

Z _,

0.00 0.00 Z

TOTAL

558

100.00

253

100.00

274

100,00

1085

100.00

8 Gq

AVG:

22

16

t6

Note: F - strictty funds-related activities L - strictly loans-related activities PC,ode* - see personnel classification code for description (Appenix 2).

Source: _mpara_ve

Bank Study,

1987,

52

JOURNAL

OF PHILIPPINE

DEVELOPMENT

comesfrom lending operations. RBs, on the other hand, have a greater bulk of their transaction cost in lending, 47.8 percent, against only 38.2 percent for their funds-mobilization activities. PDBs have almost the same transaction cost for their lending operations, 38.6 percent, and deposit mobilization activities, 36.8 percent. KBs being only a part of a nationwide bank network act as collecting stations by mobilizing and raising funds for their head offices (see Relampagos [1988]). Obviously, the emphasis is to generate as much funds from the public for their head offices. In contrast, RBs,being unit banks perform a fully dual operation of funds mobilization and lending operations with emphasis on the latter.

Furthermore, RBs,more than KBs or PDBs rely more heavily on funds from the Central Bank's rediscounting window and from special lending programs. This is shown by a bigger percentageof KBs transaction cost coming from activities related to dealings with bank depositors and clients, 29.5 percent, against RBs 9.6 percent only. On the other hand, RBs have a higher percentage of their transaction cost in activities dealing with the CB, 6.7 percent, compared to KBs only 2.6 percent. PDBs like KBs, incur substantially more costs on deposit-mobilization from the public than on obtaining rediscounted funds from the Central Bank.

The above findings is further supported by the percentage of time-allocation of personnel of the different bank types between funds-mobilization and lending activities (Table 7). About 60 percent of total personnel time by KBs are devoted to funds- mobilization against only 15.6 percent for lending operations. In contrast, RBs have only 33.0 percent of total personnel time infunds-mobilization but 51.7 percent oftotal personnel time in lending operations. PDBs also have a greater portion of their personnel time allocated to funds-mobilization (40.5%) than to lending operations (33.8%).

In terms of personnel distribution (see Table 6), KBs have more of their total personnel in funds-mobilization activities, 44.2 percent, and only 7.0 percent in lending activities. But RBs have only 11.0 percent of their personnel involved in funds-mobilization activities against 23.4 percent of their personnel in lending operations.

In summary, RBs concentrate more on their lending activities compared to KBs. This is supported by RBs' personnel time allocation and distribution in favor of their lending operations. PDBs have a more balanced operations between funds-mobilization and lending operations as evidenced by their equal share in costs between these two operations.

For KBs and PDBs, their lending operations and activities are shared with the head offices to the extent that they are given only a certain level of amount of authority in lending beyond which only their regional or head offices already assume the decision. RBs are unit banks performing both funds-mobilization and lending perhaps with a strong emphasis on

c-

Table 7 PERSONNEL TIME-ALLOCATION (percentage) ALL BANKS weighted shares

%gt

TOTAL LENDING:

721782.4

KBs

z--4 > z PDBs

%,st

weighted shares

%gt

%st

weig hted shares

23.97

100

319797.8

15.57

100

t75612.0

RBs

%gt

%st

weighted shares

33.77

100

226372.5

c_cL

%gt

%st

51.74

100

Cm < Cq

-1

Planning & Programming

58014.60

1.93

8,04

35884.83

1.75

11,22

9039.119

1.74

5,15

13090165

2,99

5,78

Ads & Promo

19338.42

0,64

2,68

7661,925

0.37

2.40

5109.275

0,98

2.91

6567,229

1.50

2,90

Interview of Applicants

63890.10

2,12

8.85

28031.90

1.36

8,77

1609g,04

3,10

9.17

19759.16

4,52

8.73

Credit Investigation

77240.63

2.56

t0,70

29664.30

1,93

12.40

17723.77

3,41

10.09

19852.55

4.54

8.77

Evaluation

82447,5"t

2.74

_,1,42

42042.92

2.05

13,15

11630.92

2,24

6.62

28773.67

6.58

12.71

oz

43556.87

1.45

6,03

16341,57

0,80

5,1 t

12734,60

2.45

7,25

14480.68

3.31

6.40

0

Monitoring

. 39458,63

1.31

5.47

24621.50

1,20

7,70

7699,025

1.48

4,38

7138.100

1.63

3.15

-1

Collection

61279,38

2,03

8,49

21683,13

1.06

6.78

21774,17

4. _,9

12,40

17822.07

4.07

7,87

15.49

& Analysis

Disbursement

>_째 z _,

Record-keeping/Report111525.2

3,70

15.45

44352.52

2.16

13,87

32107,63

6,17

18.28

35065.04

8,01

Mgt, of bad debts

writing

50836183

1,69

7.04

14132.47

0,69

4,42

16372,38

3, 15

9,32

20331.97

4.65

8.98

Unspecified

114018,9

3.79

15,80

45379,00

2,21

14,19

25322,10

4.87

14,42

43317,86

9.90

19.14

INVEST M ENTS

18661.81

0,62

330,115

0,16

7678.65

1,48

7644.049

1,75

TRUST OPERATIONS

18372.64

0,61

1331 t,54

0.65

4787.55

0,92

273,5495

0,06

w

o1 4:=

Table 7, continuation ALL BANKS

KBs

weighted sh ar es

PDBs

weighted %gt

%st

s ha res

RSs

weighted %g t

%s t

sh a res

weighted %g t

%st

sh a res

%gt

%st

TOTAL FUNDSMOBILfZ.ATION Transactions with CB,

1577807.0

other banks Transactions With Depositors

108687.6

52.39

3.61

100

6.89

1222778.0

70743.54

59.53

3,44

100

5.79

210675.1

13858.79

40.51

2,66

100

6.58

144353.9

32.99

100

24085.31

5.50

16,68

737788.3

24,50

46.76

607134.7

29.56

49.65

95340.32

18.33

45,25

35313.21

8,07

24.46

Record-keeping

371284.6

12.33

23.53

290563.4

14.15

23.76

10.21

25.2()

27629,61

6.31

19.14

Funds-Transfer

79859.79

2.65

5.06

6537g.52

3.18

5.35

53091.58 11 850.43

2.28

5.62

2629.835

0.60

1

Ads & Promo Unspecified

_01347,4 178568.3

3.37 5,93

6.42 11.32

75365.4g 113565

3.67 5.53

6.16 9,29

16769.72 195119.62

3.22 3,75

7.96 9.27

9212.212 45483,71

2.11 10.39

6.38 3t,51

GENERAL ADMINISTRATION/ SERVtCES

675014.1

22,41

494773.0

24.09

121330.5

23.33

58910.52

13.44

Grand TOTAL

3011638

t00

2053991.

100

520093,0

100

437554.5

100

Note: gt - grand total st -- sub-total Source: Comparative Bank Study, 1987.

c

m 82

z _> ro "1o -r F "_ m m < m i-0 rn z

UNTALAN

;

and CUEVAS:

TRANSACTION

COST

55

the latter. This is explained by the role of RBs as conduits,and to some extent,PDBs,forthe variousspecial-lendingprogramsof the CentralBank.

Transaction

Cost of Lending

The transaction cost of lending for banksmay be broken-downinto two majorcomponents:loanprocessingcost and loanrecoverycost.Consideringthecompositionof thetotallendingcosts (see Table 8) of the 49 banks in the sample about 41.3 percent of lendingcosts come mainly from loanrecoveryeffortssuchas monitoringof loans,collection,recordkeepingand managementof bad debts. Loan processingactivitieswhich includesinterviewsof credit applicants,credit investigation,evaluation and analysis and loan disbursementaccount for 30.5 percent of total lendingcosts.

By bank types, the contributionof loan processingactivitiesto total lendingcost is 28.5 percentfor PDBs and 28.7 percent for RBs, noting almostnodifference in their loanprocessingcostsin relationto their total lendingcosts. On the otherhand,about32.7 percento! KBs'lendingcosts are accountedfor by loanprocessingactivities.Thiscontrastindicatesthat KBs devote more resourcesto loan processingbefore approval and disbursement.

It is importantto note, however,that despite RBs and PDBs having the same loanprocessingcostsrelativeto theirtotal lendingcost,there is a differencein their costs arisingfrom credit investigationand evaluation/analysisof loans. Comparedto RBs, KBs and PDBs have their loan processingcostsaccountedmoreby creditinvestigation,12.5 percentfor KBsand 13.1 percentfrom PDBs, againstonly 8.0 percentfor RBs (see Table8). On the otherhand, RBsgivemore emphasisto the evaluation/ analysisof loansthanPDBs. Assumingthat the characteristicof their borrowers may serve to explain the difference in credit investigation cost, a comparison of the number of loans granted to repeat borrowers was made. It is expected that a bank with more repeat borrowers would spend less on credit investigation cost since it is likely that the same borrowers would offer the same security.

There is also the fact that the bank already knew other important charac_ teristics of these borrowers.

Table 9 showsthat onthe average, KBs have 24repeat borrowers per year, PDBs have 61, while RBs have a far greater average of 641. Of the 38 loan applications received by a KB,24 or 63.1% were repeat borrowers.

One notes that the RBs averaged 641 repeat borrowers, this number is 62 percent of 1,023 loan applications. Given the almost similar ratio of repeat borrowers to total applications, it can be said that the difference in credit

Table 8 LENDING COSTS (in thousand ALL BANKS Pesos

pesos)

KBs %

Pesos

O_

PDBs %

Pesos

100.00

5401.50

RBs %

Pesos

%

5789.64

100.00

TOTAL LENDING COSTS: Planning & Programming Ads & Promo

20694.47

100.00

9503.33

100.00

1468.17

7.09

902.52

9.50

233.83

4.33

331.82

5.73

440.33

2.13

162.58

1.71

115.65

2.14

162.10

2.80

Disbursement

1054_60

Unspecified

2886.99

5.10 13.95

412.1 i 1128.10

4.34 11.87

286.98 677.85

5.31 12.55

355.51 I081.04

6.14 18.67 C

Loan Processing: Interview of 1646.16

7.95

693.84

7.30

446.38

8.26

505.94

8.74

z ro -n "13

Credit Investigation

2356.23

11.39

1186.04

12.48

709.32

13.13

460 87

7.96

!r-

Evaluation & Analysis

2305.38

11.14

1229.27

12.94

382.41

7.08

693.70

11.98

6307.76

30.48

3109.15

32.72

1538.11

28.48

1660.50

28.68

Applicants

"o Ill r'13

< 1"13 r

Š

"13 m

Z

C Z -H

Table

8, continuation

r Z

ALL BANKS Pesos

KBs

PDBs

RBs C

%

Pesos

%

Pesos

%

Pesos

%

t 054.95 1526.89

5.10 7.38

641.08 518.33

6.75 5.45

212.43 560.96

3.93 10.39

201.44 447.60

3.48 7.73

Record-keeping/ Report-writing

2886.19

13.95

1148.54

12.09

880.90

16.31

856.75

14.80

Z

Mgt. of bad debts

3068.59

14.83

1480.91

15.58

894.79

16.57

692.89

11.97

oC_

8536.63

41.25

3788.86

39.87

2549.08

47.19

2198.68

37.98

-_

Loan Recovery: Monitoring CoLlection

Source:

Comparative

Bank Study,

m <

__ >:째 _z >

_o

1987.

Ol

Table 9 NUMBER OF REPEAT BORROWERS ALL BANKS

KBs

_o

PDBs

RBs

REPEAT BORROWERS

Number

0 50 100

%

Number

%

Number

3 26 2

6.82 59.09 4.55

2 15 1

11.11 83.33 5.56

0 11 0

150

i

2.27

0

0.00

0

200

2

4.55

0

0.00

2

0 0 10

0.00 0.00 22.73

0 0 0

0.00 0.00 0.00

0 0 0

44

t00.00

18

100.00

13

250 300 300 + TOTAL AVG:

% 0.00 84.62 0.00

Number

%

1 0 1

7.69 0.00 7.69

0.00

1

7.69

15.38

0

0.00

0.00 0.00 0.00

0 0 10

0.00 0.00 76.92

100.00

13

100.00

24

61

641

47 2169

1t 3 12744

465 215968

C

m Z >

1-" O _

-n "o "7

r"o -o

SD: VAR." MIN:

0

MAX: Source:

1590 Comparative

Bank

Study,

1987

tl1

o

ITF < Ill

r0

rn Z -I

C

Table 10 LOAN APPLICATIONS

z > r" _>

ALL BANKS

KBs

PDBs

RBs

z CL

No, of Loans

Number

%

Number

%

Number

%

Number

% m

<

100

31

58.49

19

90.48

11

73.33

1

5.88

300

5

9.43

2

9.52

3

20.00

0

0.00

500

4

7.55

0

0.00

0

0.00

4

23.53

700

2

3.77

0

0.00

1

6.67

1

5.88

900

2

3.77

0

0.00

0

0.00

2

11.76

1100

3

5.66

0

0.00

0

0.00

3

17.65

1300

1

1.89

0

0.00

0

0.00

1

5.88

1500

1

1._

0

0.00

0

0.00

1

5.88

1500

4

7.55

0

0.00

0

0.00

4

23.53

TOTAL:

53

100.00

21

100.00

15

100.00

17

100.00

AVG: SD: VAR:

382 530 280796

38 57 3274

161 187 34922

>:0 Z 6 z ,_O

1023 507 257477 G

MIN: MAX:

4 1854 O1

Source:

Comparative

Bank Study,

1987

¢o

60

JOURNAL

OF PHILIPPINE

DEVELOPMENT

investigation cost between KBs and PDBs on one hand, and RBs on the other is not due to the frequency of repeat borrowers in the banks' portfolios.

Table 11 indicates that KBs and PDBs require higher loan- to-collateral ratios than RBs. The average loan amount is 61 percent of the collateral for KBs and 58 percent of the collateral for PDBs. RBs, on the other hand, give loan values of 57 percent, on average, of the collateral offered. The fact that KBs and PDBs have more commercial loans in their portfolio, usually of larger amounts than agricultural loans probably explain the importance of credit investigation, i.e., inspection and appraisal to ascertain the true value and authenticity of the collateral offered in these banks. As shown in Table 7, KBs and PDBs allocated a higher percentage of personnel time to credit investigation activities, 12.4 percent and 10.1 percent for KBs and PDBs, respectively, against only 8.8 percent for RBs. Part of the credit investigation cost of loan processing is also accounted for by insurance premia paid by these banks to the special lending programs.1 Table 12 shows that not a single RB has paid guarantee fees to these programs indicating that they have not participated in these programs or that they are not accredited at all. On the other hand, KBs and PDBs have paid guarantee fees from P20,000 to as high as P160,000. On average, PDBs pay P21,707of guarantee fees while KBs pay P11,608. ThJsguarantee fees contribute further to their credit investigation cost. Likewise, participation in these programs may require additional credit investigation work which would again partially contribute to the higher personnel cost in loan processing among KBs and PDBs compared to RBs. Table 8 showed that the incidence of loan recovery costs in total lending costs is slightly higher for KBs (39.9%) and much higher for PDBs (47.2%) than for RBs (38.0%). Although rural banks service more loan accounts, yet smaller in loan value, than either PDBs or KBs, the higher loan recovery costs incurred by KBs and PDBs is due to the importance of loan recovery operations of these banks due to the larger exposure by KBs and PDBs to commercial loans than agricultural loans, the former loans being larger in amount. Among bank types, KBs and PDBs incur higher risk-related costs in managing their bad debts such as default expenses, litigation and provisions for bad debts. On the average, a KB incurs P46,665 in risk_relatedcosts whereas a PDB and an RB incur about P18,682 and P12,759, respectively (see Table 13). The difference in cost may be due to the higher loan values for KBs and PDBs compared to RBs.

1.

These were mostly fees to the crop insurance program.

c-

Table 11 LOAN TO COLLATERAL RATIO

z > IZ

ALL BANKS Loan as % of Collateral

Number

0

KBs

%

Number

PDBs

%

Number

RBs

%

Number

CL O

%

c m <

1

2.33

0

0.00

1

7.14

0

0.00

-4 :;O

1

2.33

0

0.00

1

7.14

0

0.00

> Z

50째/0or tess

14

32.56

6

37.50

3

21.43

5

38.46

_ >

75% or less t00% or less

22 5

51.16 11.63

8 2

50.00 12.50

6 3

42.86 2t.43

8 0

61.54 0.00

o 5 z

100 +

0

0.00

0

0.00

0

0.00

0

0.00

o

TOTAL:

43

100.00

16

100.00

14

100.00

13

100.00

AVG: S D:

59 20

25% or less

Source:

Comparative

Bank

Study,

1987.

61 16

58 26

57 15

-4

Table 12 GUARANTEE FEES (in thousand pesos) ALL BANKS Guarantee Fees

Number

0

KBs

%

Number

43

81,13

11

20 & less

1

1.89

40 & less 60 & less

5 2

9.43 3.77

80 & less 100 & less

0 0

0.00 0.00

0 0

120 & less 140 & less

0 1

0.00 1.89

160 & less

1 53

TOTAL:

PDBs

%

Number

RBs

%

Number

%

64.71

10

71.43

22

100.00

0

0.00

1

7.14

0

0.00

4 2

23.53 11.76

1 0

7.14 0.00

0 0

0,00 0,00

0.00 0.00

0 0

0.00 0.00

0 0

0.00 0.00

0 0

0.00 0.00

0 • 1

0.00 7.14

0 0

0.00 0.00

1.89

0

0.00

1

7.14

0

0.00

100.00

17

100.00

14

100.00

22

100.00

r t--

o I

r m

AVG:

9.457

11.608

21.707

0.000

o

SO:

27.057

0.276

46.384

0.000

< m

VAR:

732.072

266.6t5

2151.444

0.000

m r

o m

Source:

Income

and

Expense • Statements,

Dec. 1986.

z -I

Table 13 RISK-RELATED EXPENSES (in thousand pesos) Provisions

for

ALL BANKS

Litigation & Bad Debts Expenses 0

Number

KBs

%

'

Number

c Z _> Z

PDBs

%

Number

RBs

%

Number

o. ¢3 C m <

%

> t._ :o _>

25

39.06

11

42.31

8

50,00

6

27.27

25 or less 50 or less

23 6

35,94 9.38

7 2

26.92 7,69

4 1

25.00 6.25

12 3

54.55 13.64

75 or less

3

4.69

1

3.85

1

6,25

1

4.55

100 or less

3

4.69

1

3.85

2

12.50

0

0,00

125 or less

0

0.00

0

0,00

0

0.00

0

0,00

150 or less

0

0.00

0

0,00

0

0.00

0

0,00

200 or less

0

0.00

0

0.00

0

0.00

0

0.00

Z

225 or less

1

1.56

1

3.85

0

0.00

0

0.00

250 or tess

2

3.13

2

7.69

0

0.00

0

0.(30

TOTAL:

64

100.00

26

100.00

16

100.00

22

AVG: SD: VAR:

27.577 55.818 3115.663

Source:

Income

and Expense

Statements,

46.665 80.511 6481,950 Dec.

1986.

18.682 31.378 984.549

_> o -1 Z

n O

t,n

100.00

12.759 15.568 242.369 O_ ¢_

64

JOURNAL

OF PHILIPPINE

DEVELOPMENT

Another possibleexplanationisthatthe higherloanrecovery costespecially for PDBsmay be dictated by the requirements of the guarantee programs for monitoring and report-writing. A higher percentageof loan recovery cost i_ attributed to record-keeping and report writing, and management of bad debts for both KBs and PDBs. PDBs incur the highest loan recovery costs relative to the total lending costs as they have more exposure to the guarantee programs among the three bank types. Further,the higher loan recovery cost among KBs and PDBs may be due to the dependence of the head offices on their branches for collection and management of loan accounts.

Transaction

Cost

of Funds-Mobilization

As shown in Table 14, a greater portion of funds-mobilization cost by all the banks comes from deposit-mobilization (47.7%) and from recordkeeping and withdrawal (23.2%). Costs accounted by activities related to transactions with CB rediscounting are only 6.4 percent of fundsmobilization cost.

By bank types, KBs' and PDBs' funds-mobilization costs are accounted mainly by deposit-mobilization activities as transactions with bank depositors. KBs' deposit-mobilization cost accounts for 52.2 percent of funds-mobilization cost compared to 43.7 percent for PDBs. RBs, on the other hand, have only 25.1 percent of funds-mobilization cost coming from deposit-mobilization. A higher percentage of KBs' and PDBs' fundsmobilization cost is also due to record-keeping and withdrawal. This is to be expected since this cost is related to the servicing of deposits by clients.

RBs, on the other hand, have a higher percentage of their fundsmobilization cost from activities related to transactions with CB rediscounting window, 17.5 percent, against KBs' 4.6 percent and PDBs' 5.3 percent. This reflects the RBs' reliance on funds from CB, and highlights the fact that this reliance is far from costless. In fact, dependence from CB rediscountingwindow may represent an important costfor the banks.

It has been shown above that a greater percentage of personnel time is allocated to deposit-mobilization activities, 49.6 percent for KBs and 45.2 percent for PDBs compared to only 24.4 percent for RBs (Table 7). A greater proportion of personnel is also assigned to depositmobilization activities by KBs and PDBs compared to RBs (Table 6).

This is explained by the greater volume of deposits serviced by both KBs and PDBs compared to RBs. Another factor is that KBs and PDBs have other accounts, such as time-deposits, to service unlike RBs. Overall, the concentration of personnel on deposit-mobilization activities by KBs and i

C

Table 14 FUNDS-MOBILIZATION (in thousand ALL BANKS

z > i-

COSTS

pesos)

KBs

> Z PDBs

RBs

= C) C

Pesos Total Transactions with CB, other banks

37010.20

% 100.00

Pesos 27241.04

% 100.00

Pesos 5146.205

% 100.00

Pesos 4622.951

% 100.00

6.31 1250.522

4.59 274.3954

5.33 810.9047

17.54

¢o _> ¢3

17636.44

47.65

Record-keeping Fund§-Transfer

Depositors

8589.628 1529.688

23.21 6443.389 4.13 1098.311

Ads&Promo

2250.033

Unspecified

4663.796

Comparative

._ > Z.

2335.823

Transactions with

Source:

> ul

Bank

Study,

6.08

14223.77

1552.757

12.60 2671.773

52.21

2250.752

43.74

1161.915

25.13

z

23.65 1293.558 4.03 355.8993

25.14 6.92

852.6812 75.47827

18.44 1.63

o ¢n

380.576,3

7.40

316.6992

6.85

9.81 586.7502

11.40

5.70

1405.272

30.40

1987.

o1 ol

66

JOURNAL

OF PHILIPPINE

DEVELOPMENT

PDBs contributeto their higher deposit-mobilizationcosts. On the other hand, a greater percentage of personnel time is allocated by RBs to transaction cost with the CB, 16.7 percent against KBs and PDBs 3.44 and 2.44 percent respectively.

Fines and penalties related to reportingrequirements with the CB and in meeting the reserve requirement contributed a greater percentage of RBs' funds-mobilization cost. On the average, this cost is P40,071 for RBs compared to KBs' P4,504 and PDBs' P6,757 (see Table 15). On the other hand, a considerable percentage of KBs and PDBs funds-mobilization cost comes trom insurance for their deposits. This is expected since KBs and PDBs have a greater volume of deposits compared to RBs. The average is P66,468for KBs, P15,514for PDBs and P12,709 for RBs (see Table 16).

Per Unit Cost ot Lending Cost Per Loan Account Outstandina Given the overall cost incurred Dy banks in their lending operations and considering the total number of outstanding loans in their portfolio, the cost per outstanding loan is about P1,380 per account (see Table 17).

This amount represents the cost per loan by all the banks combined.

Part of this cost per account comes from processing the new loans granted for the period considered and a bigger par{ comes from servicing these new loans in addition to other loans that are already outstanding.

By type of bank, RBs have the lowest cost per loan account, (P473) than PDBs (P1,839) and KBs (P14,500) (see Table 17). The big difference in cost per loan between KBs and RBs is that not only do KBs incur higher cost in their lending operations but that they have less accounts to service. In contrast, not only do RBs incur less total costs in the lending operations but they also service more accounts. This is typical among rural banks where most loans in their port-folioare small, but numerous. PDBs also have less number of accounts in their portfolio than RBs although greater than KBs. Most of the loans by RBs are agricultural loans compared to KBs which have predominantly commercial loans.

Recovery cost associated with all outstanding loans,2 is P 564 per account for all of the banks. For RBs the loan recovery cost per account is P 166 against PDBs' P 772 and KBs' P 6,305. In all aspects of loan

2,

Total loan recovery cost divided by the total number of loans outstanding,

c

Table 15 DEPOSIT-RELATED EXPENSES

z --t _>

(in thousand pesos) ALL BANKS Fines & Penalties 0

KBs

_, z PDBs

RBs

c_ c m <

Number

%

Number

%

Number

%

Number

%

>_

24

43.64

10

52.63

11

78.57

3

13.64

:o _>

20 & less

20

36.36

7

36.84

3

21.43

10

45.45

_z

40 & less 60 & tess

2 5

3.64 9.09

1 1

5.26 5.26

0 0

0.00 0.00

1 4

4.55 18.18

5 z

80 & less 100 & less

1 0

1.82 0.00

0 0

0.00 0.00

0 0

0.00 0.00

1 0

4.55 0.00

100 +

3

5.45

0

0.00

0

0.00

3

13.64

55

100.00

19

100.00

14

100.00

22

100.00

TOTAL AVG: SD: VAR'

17.849 40.676 1654.575

M1N: MAX:

0 188.6

Source: Income and Expense Statements, Dec. 1986.

4.504 11.554 133.500

6.757 16.767 281.117

40.071 59.125 3495.821

ol

Table 16 DEPOSIT-RELATED EXPENSES (in thousand pesos) ALL BANKS Deposit Insurance 0

Number

KBs

%

Number

PDBs

%

Number

RBs

%

Number

%

1

1.85

0

0.00

1

7.14

0

0.00

20 & less

32

59.26

3

16.67

9

64.29

20

90.91

40 & less

10

18.52

4

22.22

4

28.57

2

9.09

60 & less

2

3.70

2

11.11

0

0.00

0

0.00

80 & less 100 & less

4 2

7.41 3.70

4 2

22.22 1 i .11

0 0

0.00 0.00

0 0

0.00 0.00

100 +

3

5.56

3

16.67

0

0.00

0

0.00

TOTAL

54

100.00

18

100.00

14

100.00

22

100.00

c r'-I1 o "1o "lr" "o

AVG: SD: VAR:

31.356 36.826 1356.141

66.468 45.490 2069.378

15.514 11.216 125.803

12.709 6.277 39.399

-o m o m < ITI r-

Source:income andExpenseStatements,Dec.1986

O m z .-I

UNTALAN

and (3UEVAS:

TRANSACTION

COST

69

Table 17 COST PER OUTSTANDING LOAN ACCOUNT (in pesos) ALL BANKS

KBs

PDBs

RBs

TOTAL LENDING COST:

1379.92

14500.28

1839.24

473.04

Planning & Programming Ads & Promo

93.96 30.54

1136,09 201.34

99.79 58.71

29.02 13.59

Interview of Applicants

94.76

777.69

153.48

39,50

Credit Investigation

172.27

1894.99

291.96

39.45

Evaluation & Analysis Disbursement

152.76 72.78

1821.55 681.54

81,36 123.60

67.47 23.88

Unspecified

198.56

1681.87

258.66

94.14

Monitoring Collection

67.92 107,42

986.29 937.21

78.41 197.00

9.,46 36.04

Record-keeping/ Report-writing

185.73

1724.72

328.37

58,80

Mgt. of bad debts

203.23

2656.98

167,91

61.69

564.29

6305.20

771.69

165.99

Loan Recovery:

Source: Comparative Bank Study, 1987.

70

JOURNAL

OF PHILIPPINE

DEVELOPMENT

recovery operations, i.e., monitoring, collection, record-keeping and management of bad debts, PDBs and KBs incur more costs than RBs (see Table 17). For example, the cost of monitoring each account is P78 for PDBs and P986for KBs against P9.41 for RBs. It must be recalledthat KBs and PDBs put more emphasis on their loan recovery operations due to greater exposure as a result of the larger commercial loans they make.

Further,PDBs and KBs participate in the guarantee programswhereas RBs do not. The difference in their loan servicing cost per account may be due to the importance of 10anrecovery and the requirements of these guarantee programs for supervision and stricter management of these accounts compared to regular accounts.

Table 18 reports the average cost per loan granted during the year.3 It is shown that PDBs and KBs have higher processing cost per loan at P1,023 and P6,745 respectively, against RBs' P120. As in the banks' loan recovery cost, all aspects of loan processing cost from screening to credit investigation and loan evaluation is higher among PDBs and KBs compared to RBs. An example would be the credit investigation cost per account for loan processing activities than RBs. This may be partly due to the need for extensive credit investigation and partly due to the participation of PDBs and KBs in guarantee programs.

Cost Per Peso Lent As regards the cost per peso of loan granted and loans outstanding for these banks, a totally different picture emerges. Considering all the banks, the cost per peso loan outstanding is P0.03 (see Table 19). This means that the cost of maintaining each peso of loan outstanding is about three centavos. For each bank type, this cost is P0.06 for RBs, P0.03 for PDBs and P0.02 for KBs. Overall, KBs and PDBs have the comparative advantage in lending compared to RBs as they are able to keep a lower cost per peso of loan they keep in their portfolios. This is a direct effect of the larger amounts of outstanding loans, in KBs' and PDBs' portfolio than in RBs. What PDBs and KBs lack in the number of loan accounts, they make it up by a higher loan amount pe.raccount.

The cost of recovering each peso of loan outstanding4 for each bank is P0.023 for RBs, P0.014 for PDBs and very negligible for KBs, P0.008 (see Table 19). Again, the slightly lower loan recovery cost per peso for

3.

Total loan processingcost

divided by total number of loans granted.

4,

Total loan recovery cost divided by the total value of loans outstanding.

UNTALANand

CUEVAS:

TRANSACTION

COST

71

Table 18 COST PER LOAN (in pesos) ALL BANKS TOTAL LENDING COST:

KBs

1237.30 20176.99

PDBs

RBs

3593.81

391.02

Planning & Programming Ads & Promo

71.87 25.62

1355.49 283.70

155.57 76.95

23.19 11.90

Disbursement

68.88

998.83

190.94

26.60

Monitoring Collection

64.87 95.79

1526.13 1238.90

141.34 373.23

11.59 29.30

Record-keeping/ Report-writing

178,41. 2540.00

586.09

59.90

Mgt. of bad debts

192.35

3440_20

595.34

47.36

Unspecified

152.75

2049.01

451.00

60.92

Interview of Applicants

103.83

1583.50

296.99

36.68

Credit Investigation

142,88

2543.67

471.93

32.15

Evaluation & Analysis

140.04

2617_58

254.43

51.44

386.75

6744.74

1023.36

120.27

Loan Processing:

Source: Comparative

Bank Study, 1987.

72

JOURNAL

OF PHILIPPINE

DEVELOPMENT

Table 19 COST PER PESOS LOAN OUTSTANDING (in pesos) i.i

I

ALL BANKS

KBs

PDBs

RBs

TOTAL LENDING COST:

0,026

0.018

0.030

0.060

Planning & Programming

0.002

0.001

0.001

0,003

Ads & Promo

0,001

0.001

0.002

Interview of Applicants

0.000

' 0.002

0.001

0.002

â&#x20AC;˘0.005

0,003

0,002

0.004

0.005

Evaluation & Analysis

0,003

0,002

0.002

0.007

Disbursement

0,001

0,001

0,002

0,004

Unspecified

0.004

0.002

0.004

0.011

Monitoring

0,001

0.001

0.001

0.002

Collection

0,002

0,001

0,003

0.005

Record-keeping/ Report-writing

0,004

0.002

0,005

0.009

Mgt. of bad debts

0.004

0.003

0.005

0.007

0,011

0.008

0,014

0,023

Credit Investigation

Loan Recovery:

Source: Comparative

Bank Study, 1987,

UNTALAN

and CUEVAS:

TRANSACTION

COST

73

KBs and PDBs compared to RBs is dictated by the bigger volume of loansthey service.

5 As regardsthecostof grantingperpesoof loan, thedifferencesacross banksdo notappear significant.This amountsto P0.013 for RBs, P0.015 forPDBsandP0.015 for KBs(seeTable20). Amongthe samplebanks,RBs grantedmoreloansin valuethan PDBs giventheir respectivecostswhich explainsthe slightlylower per peso costof grantinga loan. On the other hand,slightlyhighercostper pesoof grantinga loanof KBs comparedto RBs is due to the fact that althoughKBs granteda highertotal value of loansthan RBs,KBsincurredhighercostof loan processingcomparedto RBs. Overallthey do not differ in their costper peso lent. This is an importantfinding,sinceit suggeststhat current RB operationsare of similar efficiency,measuredby costsperpeso lent,comparedto KBsand PDBs.

Per

Unit Cost

of Deposit-Mobilization

Cost Per Deposit Account Consideringall the banks, their overall cost of mobilizing each depositaccount,i.e., openingof new accountsto servicingeach account, is P87 (see Table21).

Mostof the costin mobilizingeach depositaccountfrom the publicis due to activitiesdirectly related to transactions with bank clients or depositors,amountingto P52 per depositaccount.Likewise,this depositmobilizationcost is largely accountedfor by record-keepingand withdrawal.

Bytype of bank,the cost of mobilizingeach depositaccountis higher for KBsand PDBs (P120 and P73, respectively)compared to RBs (P29).

The higher cost per deposit for KBs comes from their higher cost in deposit-mobilization relativeto the numberof depositsattracted.

Muchof thiscostof mobilizingdepositaccountsfor all thethree bank typescome from activitiesrelatedto servicing new depositorsor clients and to keepingeach depositor'saccountwith the bank. Servicing each bank depositorincludesthe openingof new accountsby new clientsto over-the-countertransactionswithdepositors,i.e., withdrawal. Maintaining each accountinvolvesrecord-keeping.For all banks, KBs have an over- the-counter transactionscost of P76 per account and a recordkeepingcost of P30. PDBs have the second highestcost with P37 and

5.

Total loan processing divided by the total value of loan granted,

74

JOURNAL

OF PHILIPPINE

DEVELOPMENT

Table 20 COST PER PESO LENT (in pesos) ALL BANKS

KBs

PDBs

RBs

TOTAL LENDING COST:

0.046

0.043

0.053

0.043

Planning & Programming

0.003

0.003

0.002

0,003

Ads & Promo

0.001

0,001

0,001

0.001

Disbursement

0.003

0.002

0,003

0.003

Monitoring

0.002

0.003

0.002

0.001

Collection

0.004

0.003

0.006

0.003

Record-keeping/ Report-writing

0.007

0,005

0,009

0.007

Mgt, of bad debts

0.007

0.007

0.009

0,005

Unspecified

0.006

0.004

0,007

0,007

Interview of Applicants

0.004

0.003

0,004

0.004

Credit Investigation

0.005

0.005

0,007

0.004

Evaluation & Analysis

0.005

0.006

0.004

0.006

0.014

0,015

0,015

0.013

Loan Processing:

Source: Comparative

Bank Study, 1987

UNTALANandCUEVAS:

TRANSACTION

COST

75

Table 21 COST PER DEPOSIT ACCOUNT (in pesos) i.

s ._.

ALL BANKS

KBs

PDBs

RBs

Total DepositMobilization Cost

87.47

120.41

73.21

28.94

Transactions with Depositors

52.25

75.63

36.71

13.97

Record -keeping

23.49

30.16

22,92

10.25

Funds-Transfer

4.49

5.65

6.50

0.91

Ads & Promo

7.23

8.97

7.01

3.81

Source: Comparative

Bank Study, 1987,

Table 22 COST PER PESO DEPOSIT (in pesos) ALL BANKS

KBs

PDBs

RBs

Total DepositMobilization Cost

0.021

0.018

0.023

0,035

Transactions with Depositors

0.012

0.011

0,011

0.017

Record-keeping

0.006

0.005

0.008

0.012

Funds-Transfer

0.001

0.001

0.002

0.001

Ads & Promo

0.002

0.001

0.002

0,005

Source: Comparative

Bank Study, 1987.

76

JOURNAL

OF PHILIPPINE

DEVELOPMENT

P23 respectively, for over-the-counter transactions with depositors and record-keeping, respectively, RBs have P14 and P10 per account for these deposit-mobilization activities.

Costs Per Peso Mobilized In contrast, the cost of mobilizing per peso of deposit is lowest for KBs (P0.018 per peso), followed by PDBs (P0.023) and RBs (P0.03S) (see Table 22). This means that for KBs, the cost of mobilizing each peso of deposit is 1.8 centavos against PDBs 2.3 centavos and RBs 3.4 centavos.

This again shows KBs' comparative advantage in raising a peso of deposit. This can be explained by the larger deposit balances per account in KBs, although they have smaller number of deposit accounts.

Normally, this is expected of KBs which are situated in more prominent locations, such as in relatively well-off communities, in addition, most KBs hold commercial accounts from businesses. PDBs, likewise, have the same advantage over RBs which have more deposit accounts than PDBs although small in value.

Summary

and Conclusion

The followingare the majorfindingsof thisstudy: 1. Funds mobilizationactivitiesaccountfor a greater part of total transactioncost among all banks than lendingoperations. KBs have a larger portion of their transactioncost contributed by funds-mobilizationthan their lending operationswhile the opposite is true for RBs. This emphasizes the fact that KB branchesare funds-generatingunitswhile RBsare more lending oriented. PDBs have a balanced operation on both fundsmobilizationand lending.

2. Consideringbanks' transactioncost on lending, KBs have a higher percentageof their lending cost accounted for by loan processingcompared to PDBs and RBs. This may be due to more intensivecredit investigationof collateraloffered among KBs. Besidesgrantingsmaller,amountsper loan, RBs are more familiar with their clientele of small borrowers having only to serve a small service area of borrowers.

Loan recovery cost also accounts for a greater share of lending cost among KBs and PDBs. This is due to intensive loan recovery efforts by these banks as a result of their higher ex-

UNTALAN

and CUEVAS:

TRANSACTION

COST

77

posure given the predominanceof commercial loans in their portfolio.

3. As regardstransactioncost on funds-mobilization,a greater part is spent on deposit-mobilizationactivities speciallyamong KBs and PDBs. On the other hand, a greater portionof RBsfundsmobilizationcost come from mobilizingfunds fromthe CB rediscounting window. This cost is shown to be a substantial component of RBsfunds-mobilizationcost.

4. The cost per outstandingloan is lowestfor RBs and highestfor KBs. Butthe cost per peso of outstandingloanis lowestfor KBs and highestfor RBs. The costof grantinga loanis lowestfor RBs thaneitherPDBs or KBs. The per pesocost of granting a loan,is alsolowestfor RBs than PDBs or KBs, althoughthe differences amongthe banksis not significant.

5. The cost of mobilizingeach peso of deposit account is higherfor KBs and PDBs compared to RBs. In contrast, KBs obtain the lowest cost of mobilizing per peso of deposit, followed by PDBs than RBs. This may again be attributed to the higher volume of deposits mobilized by both KBs and PDBs. The contrast in the composition of transaction cost among the different bank types particularly KBs and RBs serves to distinguish the direction of their operations. Being only a part of a larger branch network, KB branches serve as deposit-mobilizing units for their head offices.

Thus, this is shown by the larger portion of their transaction cost in fundsmobilization. On the other hand, RBs which are unit banks can only expect to serve a limited clientele with less incentive to raise funds from depositsbut more inclined to source funds from CB. As channel for such funds their emphasis is on lending. But despite the stark contrast of emphasisin their operations, the fact remains that KBs and PDBs with larger operations hold a comparative advantage in either funds-mobilization and lending operations measured by per peso cost of delivery.

Primarily,the problem addressed is the viability of rural financial intermediafies in terms of lower transaction cost most specificallythe per unit cost of bringing bank services to the rural sector. The fact that KBs and PDBs have relatively lower cost per peso of loan and cost per peso deposit mobilized than RBs indicatestheir comparative advantage in both funds-mobilization and lending activities. But this does not mean that smaller banks which carry predominantly agricultural loans in their portfolio need go into large scale lending in order to reduce their per peso cost. In fact, the results of a related study (see Untalan 1988) reveal that

78

JOURNAL

OF PHI LIPPINE DEVELOPMENT

agriculturallendingis not a significantdeterminantof bank transaction cost. The extentand leeway ofoperationsof banks serve as significant factorsin the deliverycostper unitfor theseservicesas evidencedby the findingsof the studyon the existenceof economiesof scale.

Additional capitalizationrequirementsfor smallerbanks especially amongunitbanks,wouldpermitthesesmallruralfinancialintermediariesto expand their operationsand improvetheir performance and viabilityby exploitingeconomiesofscaleintheiroperations.Biggeroperatingcapacity for smallerbanks would lowertheir transactioncost and thus effectively lowertheiraveragecost of delivery. One way by whichthese bankscould increasetheircapitalbase isthe removalof thepresent25 percentlimiton capitalsubscriptions.

Liberal bank entry will, likewise,proveto be beneficialin reducing transactioncost in the long-runsince competitionwouldforce ruralintermediadesto producethese bank servicesat the lowestpossiblecost in orderto remain profitable. Perhaps the higherper unitcost among RBs maybe clueto the lack of incentivesto minimizecostsin the absenceof competition. Likewise, free bank entry would providethese banks a chanceto expand their operations. Wider operationsfor unitbanks like RBs provideadditionalincentivesfor expandedlendingin terms of the numberof loansby servingotherareas. Thesewillalsoservetostrengthen theirdeposit-mobilization activitiesand effectivelylowertheirfundsmobilizationcostwhilecorrectingtheirstatusas mereconduitsof funds. A higher volumeofdepositsofthesamecostcouldlowerthecostperpesoofdeposit.

In otherwords,banks,when providedincentivesto expandtheir operations can improvetheir performanceby taking advantage of the presence of economiesof scale.

Perhaps,the profitabilityand viabilityof ruralfinancialintermediaries can also be directly addressed by looking into factors affecting bank transactioncost. One way of loweringtransactioncost is through improvementsin farm productivity.This directly lowers the dsk faced by banks. It is common knowledge that this risk comes trom the beneficiariesof credit in this case the rural householdsin the form of lowerrepayingcapacity.

Improvementsin infrastructuresuch as farm to market roads,irrigation, availabilityof betterfarm inputsand equipment, bettereducationto farmers of modemtechniquesof farming,marketingassistance,and appropriatepricingpolicieswillgo a longway in increasingfarm productivity and improvingthe incomes of rural households. These reduce risk-related costs of rural financial intermediaries,and thus theirtransaction cost.

Further,improvementof ruralhouseholdincomewould=monetize"an otherwisedormant sectorof the economythusgivingincentivesfor these

UNTALAN

and CUEVAS:

TRANSACTION

COST

79

householdsto seek morecreditwhich can be translated not only in terms of the increased number of loans by the banks but an increase in the size of the loan as well. Both have decreasing effects in the per unit and per peso cost of delivery for these rural financial intermediaries.

UNTALAN

and CUEVAS:

TRANSACTION

COST

Appendix

81

1

A. LENDING OPERATIONS: 1. Planningand Programming(e.g. setting-upof loantargets/ programs) 2. Advertisingand Promotions 3. Interviewof CreditApplicants/Examination of Loan Applications 4. Credit Investigation(e.g. inspection/appraisal of collateral; examinationof bank) 5. Evaluation/Analysisand Approvalof Loans 6. Disbursementof Loan 7. Monitoringof LoansIncludingTechnicalAssistance" 8. Collectionof Loans 9. Record-keepingand Report-writing 10. Managementof Bad Debts B. INVESTMENTS C. TRUST OPERATIONS D. FUNDS-MOBILIZATION: 1.

2.

3.

4.

5.

Transactionswiththe CentralBank/otherbanks TransactionswithBank Depositors Record-keepingand Withdrawal Funds-TransferOperations Advertisingand Promotions

E. GENERALADMINISTRATION/SERVICES (e.g. typing,delivery/messengerialactivities, maintenance/utility)

JOURNAL

OF PHILIPPINE

DEVELOPMENT

Appendix 2 Personnel Classification Code

Code 1 2 3 4 5 6 7 8 9 10

11 12

13 14

15

16 17

Bank PersonnelPositions Chairman Vice-Chairman Directors/President Board Secretary/Treasurer Manager/President-Manager AssistantManager/BranchOperationsManager/ Branch Operations Officer Cashier/Assistant Cashier/Cash Clerk BranchAccountant/Accountant/General Bookkeeper/Assistant General Bookkeeper Loans officer/Account Officer/Credit Administrator Senior Teller/Head Teller/GeneralTeller/PR Teller/Field Teller New Accounts/ Savings Pro/Customer Relations Assistant FX Clerk/CTD Clerk/Sundries Clerk CA Bookkeeper/SA Bookkeeper/ Supervising Bookkeeper/Junior or Senior Bookke.eper/LiabiUtyBookkeeper/ CTD Bookkeeper/Posting Clerk/ Proofsheet/Accounting Clerk (Funds) Clearing Clerk/BatchingClerk/ Distributing Clerk Loans Analyst/Loan Processor/ Loan Clerk/Credit Investigator/ CreditAnalyst/FinancialAnalyst/ Clerk/Loans-Rediscount Clerk Loansbookkeeper/Accounting Clerk(Loans)/Subsidiary-ledger Bookkeeper/Filing Clerk Inspector/Technician/Farm or Credit Technician/Production Technician Settling Clerk/Branch Courier/Messenger/ Utility .Clerk

UNTALANand

18 19 20 21 22 23 24 25

CUEVAS;

TRANSACTION

COST

â&#x20AC;˘ Secretary/ClerkTypist Driver/Janitor/Messenger/Security Guards Appraiser/CostingClerk Collector Property-LiaisonClerk AcquiredAssetAdministrator AdministrativeAssistant/PersonnelPro MoneyShopManager/MoneyShopSupervisor

Appendix

3

A. Where the personnel/positiondiffers by name but having more or less similar functions, these are grouped together as one classification and assigned one variable. ex. 1

Senior Teller, Head Teller, General Teller

ex. 2

Loans Analyst, Loans processor. Loans Clerk

B. Where the personnel/position differe slightly in functions but can be categorized as one general office function or activity these are assigned one variable, i.e. deposit-taking, these are grouped together in one classification, ex. 1.

Senior Teller, Field Teller, New Account Clerk, Savings Personnel

ex. 2.

inspector, Farm Technician, Credit Technician, Production Technician

ex 3.

Current Account Bookkeeper, Savings Account Bookkeeper, Certificate of Time Deposit Bookkeeper, Posting Clerk (Savings)

C. For personnel/positionsthat belong to the same classification as to deposit-taking or lending, but differ in rank, i.e. officer-positions vs. rank and file, these are assigned one variable.

83

84

JOURNAL

ex.

OF PHI LIPPINE DEVELOPMENT

Branch Accountant, Accountant, General Bookkeeper, Assistant General Bookkeeper (8) vs. C.A. Bookkeeper, S.A. Bookkeeper, Supervising Bookkeeper, Jr. and Sr. Bookkeeper, Accounting Clerk (Deposit).

D. For positionsthat have general descriptions but involvingcompletely different office function on activity, segregation is made by noting the % of their time devoted to the major functions i.e. deposits or lending. ex

Accounting Clerk (Deposit-taking) Accounting Clerk (lending operations)

E. Other positions which are distinctly attributed to a particular bank are assigned separate variables to avoid arbitrary lumping or classification. ex.

PCIB Money Shop Manager/PCIB Money Shop Supervisor

The above insures that the grouping of personnel performing similar or slightly different office activities belong to the same major office functions activity as required in the time allocation table (A) Lending, (B) Investment (C) Deposit-taking,(E) General Administrative. The above guidelines were based on job descriptions of each personnel and/or by noting the amount of time allocated to each of the major office function, (A) - (E).

Majorityof the banks surveyed do not have a complete matching of time-allocation of each personnel against the corresponding compensation of such personnel. In order that whatever existing information on these banks can be used, valuesfor these missing data were generated and the following guidelines were followed:

A. Positionswithnocompensationbutwith time-allocation .1..RBs --averaging all compensationfor that particularpositions acrossall RBsand takingintoconsiderationthatthe resulting compensationiswithinthe salaryrange forthe bank in question i.e. the computed compensationfor teller of RB1 must

UNTALANandCUEVAS:

TRANSACTION

COST

85

not be higher than the compensation of the cashier of the same bank. Otherwise, re-calculation is made by deleting the highest compensation value in the sample until the computed compensation is within RB1 salary range.

2. KBs/PDBs r averaging all compensation for that particular position using only existing values of branches of that bank under consideration i.e. teller position BPI San Pablo generated using compensation of other teller positions of other BPI branches.

B. Positionswith no time-allocation but with compensation 1. RBs -- averaging time-allocation of that particular positions across all RBs, i.e. time allocation of manager is computed by averaging all time allocation for managers by all RBs. 2. KBs/PDBs-- averaging all time-allocation for that particular position using time-allocation of personnel from other branches of the same bank.

C. Officer Positions with no time-allocation 1. RBs -- for positions of Chairman, Vice- Chairman, Board Members that have no time allocation,values are given using equal time allocation of 50 per(_entfor lending and 50 percent for deposit- mobilization.

UNTALANandCUEVAS:

TRANSACTION

COST

87

REFERENCES Cuevas, Carlos E. "Costs of Financial Intermediation Under Regulations: Commercial and Development Banks," September 1984.

Mimeographed. and Douglas Graham. "The Importance of Transaction Costs in Domestic Resource Mobilization and Credit Allocation," April 1984. Mimeographed.

Graham, Douglas. "Notes on Comparative Bank Study," December 1987. Unpublished.

Lamberte, Mario B. "Comparative Bank Study: A Background Paper."

PIDS Working Paper Series No. 87-04. Makati: Philippine Institute for Development Studies, April 1987. and Joseph Y. Lim. "Rural Financial Markets: A Review of Literature." PIDS Staff Paper Series No.87-02. Makati: Philippine Institute for Development Studies, January 1987.

Relampagos, Julius. "Funds Transfer Operations and Savings Mobilization Among Branch Banks in the Rural Areas." Unpublished M.A. thesis. University of the Philippines' School of Economics, 1988.

Tolentino, V. Bruce J. "Current Imperatives and Development in Philippine Agricultural Credit Policy," March 1987. Unpublished.

Untalan, Teodoro S. "Transaction Cost of Banks: A Comparative Study."

Unpublished M.A. thesis. University of the Philippines' School of Economics, 1988.